Sequential pattern extracting apparatus

Abstract

Constraining sequential data expressing sequential data which a sequential pattern to be extracted must include is specified in advance. Sequential pattern candidates with sequence length 1 are initially determined from among a plurality of input sequential data. Next, a set of sequential pattern candidates is generated by determining a plurality of new sequential pattern candidates by elongating the sequence length of the sequential pattern candidates. In this sequential pattern candidate set, after sequential pattern candidates which can generate only a sequential pattern which does not include constraining sequential data are eliminated, sequential data which include constraining data, and frequently appear are extracted as a new sequential pattern.

1 - 7 . (canceled) 
   
   
       8 . A sequential pattern extracting apparatus which extracts a sequential pattern frequently appearing in a plurality of sequential data each including items, comprising:
 a first storage device to store constraining sequential data expressing sequential data which a sequential pattern to be extracted must include;   a decomposition device configured to decompose one sequential data among the plurality of sequential data into first partial-sequential data and second partial-sequential data with a position at which one of the items of the constraining sequential data appears being as a boundary;   a first extracting device configured to extract a first partial-sequential pattern including the constraining sequential data from the first partial-sequential data, and to extract a second partial-sequential pattern including the constraining sequential data from the second partial-sequential data;   a merging device configured to merge the constraining sequential data, the first partial-sequential pattern, and the second partial-sequential pattern, thereby to obtain merged sequential data; and   a second extracting device configured to extract the sequential pattern including the merged sequential data frequently appearing in the plurality of sequential data.   
   
   
       9 - 15 . (canceled) 
   
   
       16 . A method for operating a sequential pattern extracting apparatus which extracts a sequential pattern frequently appearing in a plurality of sequential data each including items, comprising:
 storing constraining sequential data expressing sequential data which a sequential pattern to be extracted must include by a first storage device;   decomposing one sequential data among the plurality of sequential data into first partial-sequential data and second partial-sequential data with a position at which one of the items of the constraining sequential data appears being as a boundary by a decomposition device;   extracting a first partial-sequential pattern including the constraining sequential data from the first partial-sequential data, and of extracting a second partial-sequential pattern including the constraining sequential data from the second partial-sequential data, by a first extracting device;   merging the constraining sequential data, the first partial-sequential pattern, and the second partial-sequential pattern, thereby obtaining merged sequential data by a merging device; and   extracting the sequential pattern including merged sequential data frequently appearing in the plurality of sequential data by a second extracting device.   
   
   
       17 . (canceled) 
   
   
       18 . A program stored in a computer-readable medium for extracting a sequential pattern frequently appearing in a plurality of sequential data each including items, the program comprising:
 means for instructing a computer to store constraining sequential data expressing sequential data which a sequential pattern to be extracted must include in a first storage device;   means for instructing the computer to decompose one sequential data among the plurality of sequential data into first partial-sequential data and second partial-sequential data with a position at which one of the items of the constraining sequential data appears being as a boundary;   means for instructing the computer to extract a first partial-sequential pattern including the constraining sequential data from the first partial-sequential data, and to extract a second partial-sequential pattern including the constraining sequential data from the second partial-sequential data;   means for instructing the computer to determine merged sequential data by merging the constraining sequential data, the first partial-sequential pattern, and the second partial-sequential pattern; and   means for instructing the computer to extract the sequential pattern including merged sequential data frequently appearing in the plurality of sequential data.